Team_1 - Checkpoint_2.0
Introduction
The analysis within this report is intended to inform and better equip the marketing manager of an alcohol distributor, operating in Iowa. An analysis of per capita sales across regions of Iowa will demonstrate where consumption of alcohol is highest, differentiated between different alcohol categories. This will allow the marketing manager to better allocate marketing material to regions where they can be most effective in increasing sales.
Data Summary
Demographic Analysis: 
The data analyzed in this report contains sales data from 2012 to 2016, measuring both dollar sales and volume sales, and demographic data for all regions of Iowa. (General ACS and Sales data analysis and visualization here)
General Sales analysis of County Level Across Categories
General Sales analysis of City Level Across Categories
Across City and County level, the sales data projected similar pattern as Whisky, Vodka and Rum dominated the most part of the sales, followed by Misc, Tequila, Brandy, Schnapps, Gin, etc. The rankings across the top 10 counties with most sales in dollars and most sales in volume stayed constant, with Polk, Linn and Scott county taking the first three across dollar sale and volume sale.
City sales saw identical ranking in popularity of product categories. However, there is change of ranking between sale in dollar and sale in volume. Cities including Waterloo, Iowa city and West Des Moines swtiched places, the reason of this swtich in places could due to the fact that Vodka’s retail prices are lower but sold in greater volume, hence causing the change in ranking, besides this, the pattern of the city level analysis stayed constant with the county level.
Data Analytics
Per Capita Sales Distribution Across Geographies
Research Question 2:
Top 10 geographies for per capita consumption across every liquor category
County Level
library(dplyr) library(tidyr)
merged_data_counties <- read.csv(“merged_counties.csv”)
grouped_data_counties <- merged_data_counties %>%
group_by(county, category)
Calculate total population and sale volume by County and category
summarized_data <- grouped_data_counties %>%
summarize(total_population = sum(population),
total_sale_volume = sum(sale.volume))Calculate per capita consumption by dividing total sale volume by total population
per_capita_data <- summarized_data %>%
mutate(per_capita_consumption = total_sale_volume / total_population)
Rank geographies by per capita consumption across all categories
counties_ranked_data <- per_capita_data %>%
group_by(category, county) %>%
summarize(avg_per_capita_consumption = mean(per_capita_consumption)) %>%
arrange(category, desc(avg_per_capita_consumption)) %>%
group_by(category) %>%
slice(1:10)
View the top 10 geographies with highest per capita consumption across all categories
counties_ranked_data
Top 10 Counties per capita consumption
City level
merged_data_cities <- read.csv(“merged_cities.csv”)
grouped_data_cities <- merged_data_cities %>%
group_by(city, category)
Calculate total population and sale volume by City and category
summarized_data <- grouped_data_cities %>%
summarize(total_population = sum(population),
total_sale_volume = sum(sale.volume))
Calculate per capita consumption by dividing total sale volume by total population
per_capita_data <- summarized_data %>%
mutate(per_capita_consumption = total_sale_volume / total_population)
Rank geographies by per capita consumption across all categories
cities_ranked_data <- per_capita_data %>%
group_by(category, city) %>%
summarize(avg_per_capita_consumption = mean(per_capita_consumption)) %>%
arrange(category, desc(avg_per_capita_consumption)) %>%
group_by(category) %>%
slice(1:10)
View the top 10 cities with highest per capita consumption across all categories
cities_ranked_data
Top 10 Cities per capita consumption
Zipcode Level
merged_data_zipcodes <- read.csv(“merged_zipcodes.csv”)
grouped_data_zipcodes <- merged_data_zipcodes %>%
group_by(zipcode, category)
Calculate total population and sale volume by zipcode and category
summarized_data <- grouped_data_zipcodes %>%
summarize(total_population = sum(population),
total_sale_volume = sum(sale.volume))
Calculate per capita consumption by dividing total sale volume by total population
per_capita_data <- summarized_data %>%
mutate(per_capita_consumption = total_sale_volume / total_population)
Rank geographies by per capita consumption across all categories
zipcodes_ranked_data <- per_capita_data %>%
group_by(category, zipcode) %>%
summarize(avg_per_capita_consumption = mean(per_capita_consumption)) %>%
arrange(category, desc(avg_per_capita_consumption)) %>%
group_by(category) %>%
slice(1:10)
View the top 10 zipcodes with highest per capita consumption across all categories
zipcodes_ranked_data
Top 10 Zipcodes per capita consumption
Top 10 geographies per capita dollar sales across all categories:
To analyze the dollar sales per capita, the datasets are imported and merged: ```{r merging_datasets} cities.acs <- read.csv(“project.acs.cities.csv”) cities.sales <- read.csv(“project.sales.cities.csv”) cities <- merge(x=cities.acs, y=cities.sales, by=“city”)
counties.acs <- read.csv(“project.acs.counties.csv”) counties.sales <- read.csv(“project.sales.counties.csv”) counties <- merge(x=counties.acs, y=counties.sales, by=“county”)
zipcode.acs <- read.csv(“project.acs.zipcodes.csv”) zipcode.sales <- read.csv(“project.sales.zipcodes.csv”) zipcode <- merge(x=zipcode.acs, y=zipcode.sales, by=“zipcode”)
Next, the per capita dollar and volume sale variables are calculated and joined to the merged datasets:
```{r per_capita_variables_generation}
cities$sale.dollars.capita <- cities$sale.dollars/cities$population
cities$sale.volume.capita <- cities$sale.volume/cities$population
counties$sale.dollars.capita <- counties$sale.dollars/counties$population
counties$sale.volume.capita <- counties$sale.volume/counties$population
zipcode$sale.dollars.capita <- zipcode$sale.dollars/zipcode$population
zipcode$sale.volume.capita <- zipcode$sale.volume/zipcode$populationThen, these merged datasets are exported to three csv files for the
differing levels of geography.
{r per_capita_variables_generation} write.csv(cities, "project.cities.merged.csv") write.csv(counties, "project.countie.merged.csv") write.csv(zipcode, "project.zipcode.merged.csv")
These csv files are then opened in Tableau to visualize the per capita
data. In Tableau, for all three levels of geography, the per capita
dollar sales are mapped, with their size and color gradient, indicating
how large sales per capita are in that region. A top ten list is added,
with the visual color breakdown highlighting the makeup of different
alcohol categories in that region. Finally, a filter is added to examine
the per capita dollar sales of specific alcohol categories and the
corresponding top ten regions. These results are combined in a
dashboard, which can be seen for all three levels of geography below
Top 10 Cities:
Top 10 Cities per capita
Top 10 Counties:
Top 10 Counties per capita
Top 10 Zipcodes:
Top 10 Zipcodes per capita
Analysis of the top regions for per capita sales:
The intuitive choice regarding where to market alcohol in Iowa may be to choose areas with the highest density, given that these areas have the largest potential customer base. However, as seen from the analysis above, the areas with the highest population density do not necessarily correlate with the highest per capita sales of alcohol. For example, Des Moines, Cedar Rapids, and Davenport are the three largest cities in Iowa. However, they are not in the top ten cities for alcohol sales per capita. High per capita sales indicate a substantial demand for alcohol in a region. This fact makes the regions highlighted in the visualizations above as potentially good targets for increased marketing. Furthermore, the analysis can be drilled down to the alcohol category level, and regions with the highest per capita sales in a particular category can be identified and marketed to, specifically for that alcohol type.
A Notable Outlier:
When examining the top cities for every alcohol category, a particular city stands out for sales of distilled spirits. In Mount Vernon, there is a much higher sale of distilled spirits per capita than in any other city in the state, and by some margin. This indicates a high demand for distilled spirits in Mount Vernon and a potentially good target for marketing our distilled spirit brand. This is demonstrated in the screenshot below:
Mount Vernon - An outlier for Distilled Spirits
Conclusion
Q2: The most popular liquor category remained constant throughout the county, city and zipcode level, the geographies with most per capita consumption also aligns with the general sales analysis data, supporting the assumption of the popularity of Whisky, Vodka and Rum. Out of all the geographies, Dickinson County and Bevington city stood out, as they have the highest per capita consumption in their geographical category respectively, while not being in the top 10 of overall sale in dollars and volume. This could mean that the drinking culture in these two places is more profound comparing to other counties and cities in the state of Iowa.
Policy Recommendation
(in the works)